The flow of safe and effective new drugs to the market has slowed in the last several years, and threatens our ability to continue to improve the nation's health. At the same time, our ability to measure biological systems and understand their function is exploding. Structural genomics has given us unprecedented access to the detailed three-dimensional structure of proteins, in many cases bound to small molecule drugs that modulate their function. Functional genomics measurements of cellular gene expression have given us a picture of how drugs impact the entire cell through their multiple physical interactions The opportunity now is to combine the detailed mechanistic understanding of drug binding and drug-target interaction with a systems view of drug effect, to create novel, high-confidence hypotheses about new therapeutic opportunities. In this proposal, we combine our informatics expertise in 3D structure analysis and gene expression analysis, with our medical and pharmacogenetic expertise to propose a research program to combine structural and functional data for drug repurposing-the use of previously approved and safe drugs for new indications, alone or in combination with other drugs. Our preliminary work has shown that we can detect distant binding site similarities to suggest new targets for existing drugs. It has also shown that we can associate novel sets of genes to diseases and drug-responses based on their patterns of expression. We propose to develop these together in the context of cancer and the treatment of rare orphan diseases. We present a focused plan to (1) develop novel methods for predicting the druggability of a protein, (2) create algorithms for detecting binding similarity between two pockets, (3) create filters that use expression data to find the most attractive pathways to target through single or multiple repurposed drugs, and (4) to apply our methods to propose new cancer therapies and new rare disease therapies. With success, our methods will demonstrate how the revolution in molecular biology and genomics can be harnessed to assist drug discovery-initially, in the context of repurposing, but with an eventual goal of designing entirely new small molecule therapies. PUBLIC HEALTH RELEVANCE: Despite huge advances in our ability to measure biological systems, this has not translated into the anticipated increase in new and effective drugs for poorly treated diseases. We outline a plan to combine two types of data (structural and functional genomics) to generate hypotheses about new ways to use existing drugs to treat cancer and rare diseases. Our computational methods will provide proof that we can use the fruits of genomics effectively to create new treatments for disease.